Internship · Technology

Data Science & Machine Learning Intern

Remote, 4 to 12 weeks · or Summer 2027, 10 weeks in New York, NY or Chicago, IL

About the team

The machine learning group is based in San Francisco and works with researchers in every Lenox office. It builds models on text, alternative and market data, and it holds them to a simple standard: a model earns its place only if it beats a transparent baseline out of sample, after costs, on data that was available at the time. Interns join the group from New York or Chicago and sit with the researchers who will use their work.

What you’ll do

  • Own one modeling question for the internship, agreed with your mentor at the start.
  • Prepare a point-in-time dataset, such as company filings or earnings call transcripts, and write down what each field means and when it became known.
  • Fit a simple, interpretable baseline before trying anything more complex.
  • Evaluate models with time-aware cross-validation that prevents information from the future leaking into training.
  • Measure how stable your results are across time periods, sectors and random seeds.
  • Present your results, including the experiments that failed, to researchers and engineers at the end of the internship.

What you bring

Required (you must already be able to): Python (NumPy, pandas), scikit-learn or PyTorch, probability and statistics, linear algebra, Git.

Helpful: natural language processing, time-series methods, SQL, experience with large datasets or GPU training.

Why you qualify

  • You are pursuing a bachelor’s, master’s or PhD in any major, and you can already do what this page lists under “Required”. Common majors for this role: computer science, statistics, mathematics, data science and related fields. Remote internships are open to students in any country; the summer program is for students graduating between December 2027 and June 2028.
  • You have trained and evaluated a model on real data, and you can say what you would do differently.
  • You understand overfitting, data leakage and why a high in-sample score proves little.
  • You prefer a result you can explain to one you cannot.
  • You do not need prior finance experience.

What you’ll learn

  • Why financial data is noisy, non-stationary and easy to misuse, and how researchers guard against that.
  • How to design an evaluation that a skeptical reviewer will accept.
  • How a research prototype becomes a monitored production model.

How the internship runs

Every internship moves through the same five stages, whether it is remote (4, 6, 8 or 12 weeks) or the ten-week summer internship.

Stage 1Onboarding, data tour, agree your question
Stage 2Dataset preparation and baseline
Stage 3Model experiments and peer review
Stage 4Stability and robustness tests
Stage 5Write-up and final presentation

How we hire

Our process is thorough. For this role it typically combines a review of your application, a short online coding and statistics assessment, a conversation about a model you have built, and a practical modeling exercise. Your recruiter will confirm each step, and we aim to reply within a week of every conversation. Read more about how we hire.

Terms

Remote internships: the length, dates and terms are stated in your offer letter. Summer internship: paid at $2,900 per week, with housing support for interns relocating more than 50 miles. Students who need U.S. work authorization may intern under Curricular Practical Training where their school authorizes it. Strong performers may receive an offer for a full-time role; an internship does not guarantee future employment. Lenox Asset Management is an equal opportunity employer.